Posted by Jamal Spelling on January 6, 2008, at 17:04:47
In reply to Re: correlation -- justifies further investigation » Jamal Spelling, posted by SLS on January 6, 2008, at 16:32:59
> > regression to the mean.
>
> What is "regression"?
>
> I am not, unfortunately, a student of statistics. I dropped out of school because my cognition and memory became so badly impaired due to a worsening of my depressive illness.Your cognition and memory seem pretty sharp to me. I'm not sure why you keep putting yourself down like that. Anyway...
I know about as much statistics as anyone. "Regression to the mean" is a phrase I picked up on Ben Goldacre's "Bad Science" column, LOL! My understanding is that, in this context, it refers to the phenomenon in medicine whereby people get better on their own (a non-depressed state being the mean).
For example, pain is cyclical. It gets worse, and then on its own, it gets slightly better, and then slightly worse, etc. So people are most likely to treat pain when it is at its worst. But this is precisely the time when it partially remits on its own, so this remission is not necessarily as a result of the pain treatment. And that is the reason why there is such a high placebo response with pain treatment, and so many alternative pain treatments which probably do not work, but still have many satisfied customers.
Anyway, I have made my thoughts on STAR*D known. People may weigh the evidence and choose to agree or disagree with me.
poster:Jamal Spelling
thread:804126
URL: http://www.dr-bob.org/babble/20080105/msgs/804673.html